ArduPilot / ardupilot

ArduPlane, ArduCopter, ArduRover, ArduSub source
http://ardupilot.org/
GNU General Public License v3.0
10.81k stars 17.26k forks source link

Copter: 4.0 issues list #12547

Open rmackay9 opened 4 years ago

rmackay9 commented 4 years ago

This is a list of issues discovered during Copter-4.0 testing:

Common support problems:

Regressions from 3.6.x:

New for 4.0.4:

Other issues:

Resolved in Copter-4.0.4:

Resolved in Copter-4.0.3:

Resolved in Copter-4.0.2:

Resolved in Copter-4.0.1:

Resolved in Copter-4.0.0-rc5:

Resolved in Copter-4.0.0-rc3:

Resolved in Copter-4.0.0-rc2:

Could not be verified OR not a problem:

tridge commented 4 years ago

fix for #12542 here: https://github.com/ArduPilot/ardupilot/pull/12677

tridge commented 4 years ago

fix for sparky2 here: https://github.com/ArduPilot/ardupilot/pull/12678

andyp1per commented 4 years ago

It's not true to say dshot600 isn't working - it's working for me fine, must be hardware specific.

GlidLov commented 4 years ago

today i tried a large quad with cube 2.1, copter 4.0 stable and a Drotek Sirius F9P + RM3100 compass. Same quad had here2 (m8p) and 3.6.9 previously working ok. Well there's something weird about compass. although it's correctly calibrated, it show ground mag anomoalies while still on ground (sometimes, not even always, let say 4 times out of 5) and does a "ground mag anomaly, inflight yaw allignment" just after take off. Sometime it happened with M8P in the past also, but it was a more consistent and predictable way, and way less often. Just a little toilet bowl, then message of the inflight yaw realignment and it was ok. The thing is that with the new setup, it does toilet bowl heavily when in loiter, then i revert back to stabilize, back in loiter, and then much less toilet bowl, then again stabilize and loiter, and then it "clicks in" and does a proper loiter. Sometime it takes four of that cycles to click in. Oncle "clicked", it will do OK for the rest of the flight. Then Land, cycle power, and again heavy toilet bowl, ground mag anomaly, etc.... I think it's something with the RM3100 driver. Satellites are doing ok, i even reached a personal record of 32 sat / 0.4 hdop fix with that!

rmackay9 commented 4 years ago

@GlidLov, thanks for the report. Could you raise a new topic in the Copter-4.0 forum and if possible include an onboard log file? This will make the analysis and communication re this report much easier, thanks!

P.S. I suspect it's interference on the compass from the motors/ESC/powerwires but it will become clear with an onboard log.

GlidLov commented 4 years ago

@GlidLov, thanks for the report. Could you raise a new topic in the Copter-4.0 forum and if possible include an onboard log file? This will make the analysis and communication re this report much easier, thanks!

P.S. I suspect it's interference on the compass from the motors/ESC/powerwires but it will become clear with an onboard log.

well, i just discovered a potential mistake (still investigating): i suspect i'm not using the RM3100 but the infamous internal compass of the CUBE. This is because i have sirius gps plugged into Telem2 (not working in the GPS2 port) which does not have an I2C port. i will have a look into logs later today.

lukezhqin commented 4 years ago

@rmackay9 I really hope you guys would kindly review this issue #13251 which I've already open this PR #13102 though a long time ago. Many thanks. I wish this wp_nav issue would be fine solved before our offical version 4.0 for Copter released although my PR may seems to be so rude and not yet enough elegant but which did work well.

anbello commented 4 years ago

I added a comment on https://github.com/ArduPilot/ardupilot/issues/12991 (EKF is not using RangeFinder when EK2_RNG_USE_HGT > 0, EK2_ALT_SOURCE = 0) I think it may be of interest.

omegaretro commented 4 years ago

I compiled a 4.0.4 master for the pixhawk1-1m, but the rm3100 is not being listed in the compass_typemask which may have been canceled during the compilation appears the rm3100 being compiled

andyp1per commented 4 years ago

Acro is broken in 4.0.4 from https://github.com/ArduPilot/ardupilot/commit/7fe700b738902c31a8a67919be3ca18bd2323cfa there is a typo:

g.acro_rp_p = constrain_float(g.acro_rp_p, 0.0f, 1.0f); should instead be g.acro_rp_expo = constrain_float(g.acro_rp_expo, 0.0f, 1.0f);

rmackay9 commented 4 years ago

@andyp1per, thanks for finding this. I guess not many beta testers fly acro! FYI @lthall.

amilcarlucas commented 3 years ago

@rmackay9 can you please add : https://discuss.ardupilot.org/t/downloading-dataflash-logs-doesnt-work-in-with-copter-4-0-5/65620 to the list